Ground Delay Program Analytics with Behavioral Cloning and Inverse Reinforcement Learning

Abstract

Historical data are used to build two types of models that predict Ground Delay Program implementation decisions and produce insights into how and why those decisions are made. More specifically, behavioral cloning and inverse reinforcement learning models are built that predict hourly Ground Delay Program implementation at Newark Liberty International and San Francisco International airports. Data available to the models include actual and scheduled air traffic metrics and observed and forecasted weather conditions. The developed random forest models are substantially better at predicting hourly Ground Delay Program implementation for these airports than the developed inverse reinforcement learning models. However, all of the models struggle to predict the initialization and cancellation of Ground Delay Programs. The structure of the models are also investigated in order to gain insights into Ground Delay Program implementation decision making. Notably, characteristics of both types of model suggest that Ground Delay Program implementation decisions are more tactical than strategic: they are made primarily based on conditions now or conditions anticipated in only the next couple of hours.

Towards an Evolved Immersive Experience: Exploring 5G- and Beyond-Enabled Ultra-Low-Latency Communications for Augmented and Virtual Reality

“…This includes a preprocessing step, which involves generating an expert dataset through supervised training based on the possible outcome of the guided beams given that the pose of tracking (feedback) is known to the user. The action is based on learning a policy through behavior cloning [ 127 ] by mapping the multi-sensory information with the respective optimal beam based on the pre-trained dataset. The reward is then determined from the action to select the best scheduling policy for a low-latency data delivery.…”

Methodological identification of anomalies episodes in ECG streams: a systematic mapping study

“…Various state-of-the-art classification techniques for different T-wave patterns are sourced from the existing literature [58,59]. In the construction of combo pack of solution for better classification of T wave anomalies, firstly designed or identified the best possible research questions after reviewing the literature [5,[50][51][52][53]. These research questions are analyzed in a systematic way and then report the solutions to these questions in a combo pack solution (better classification and visibility of T wave anomalies).…”

“…This SLR is a joint alignment of different behaviors of different T wave episodes along with a discussion of T wave dependencies analysis. Finally, the usage of the above combination with neural models works for the achievement of robust and accurate classification of different T waves [50][51][52][53][54]. In the context of accurate and robust classification, the proposed idea is executed with the query string generation method.…”

“…The methods of DWT for T wave detection, T wave detection algorithm, TRA approach, and wavelet analysis all are worked under the coverage of parametric analysis of T wave classification. Figure 11 indicates the addressing literature in a pictorial model [3,49,50]. The novelty of behavior analytics in this combination is playing the critical role for highlighting the dependences of T wave on the other features of ECG.…”

Computational Missile Guidance: A Deep Reinforcement Learning Approach

“…To this end, it would be more beneficial to develop data-based guidance algorithms for guidance problems that suffer from uncertainties, e.g., target movement and aerodynamic force. Considering the properties of the guidance problem, leveraging the reinforcement learning (RL) concept might be most appropriate for developing a data-based guidance algorithm [15,16]. Previous works using RL to solve control problems mainly focused on the applications of robotics, with few works addressing aerospace guidance problems.…”